Files
oikos/knowledge/sources/references/cert-sync-and-traefik-config.md
dtoro 8a6422bd7d docs: move narrative wiki under knowledge/wiki/ (phase 3)
Problem: node and cross-cutting narratives lived at the repo root
(containers/, vms/, infrastructure/, host .md files), interleaved with the
machine-readable substrate.

Change:
- Move containers/ -> knowledge/wiki/containers/, vms/ -> knowledge/wiki/vms/,
  infrastructure/ -> knowledge/wiki/infrastructure/, hosts/{hubris,strong}.md ->
  knowledge/wiki/hosts/, infrastructure/references/ -> knowledge/sources/references/,
  GLOSSARY.md -> knowledge/GLOSSARY.md.
- Add knowledge/{index.md,log.md,sources/index.md} scaffolding.
- Rewrite all relative links repo-wide via a path-resolving mapper (inbound +
  outbound + between-moved-files), including .hermes/, runbooks, operations,
  investigations, plans, README, AGENTS.
- Repoint inventory.yaml doc_page fields and regenerate hosts/*.yaml (which
  embed doc_page); update oikos/gen-topology.py output path, candidate doc
  paths, and footer links; update code-comment doc paths.

Substrate untouched in place: inventory.yaml, hosts/*.yaml (regenerated,
idempotent), oikos/ code, mcp/, secrets/, bin/.

Verification:
- Logical broken-link set identical to pre-move baseline (net 128 -> 127; the
  topology regen fixed one, introduced none). Remaining are pre-existing refs
  to destroyed/archived nodes, out of scope for this move.
- gen-topology.py --check exit 0 (in sync); cards carry knowledge/wiki/ doc paths.
- build_host_files.py idempotent; all inventory doc_page targets resolve.
- MCP contract verified: get_page/search_docs/get_changelog resolve moved pages.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-06 14:35:23 +02:00

351 lines
10 KiB
Markdown

# Current cert sync script + traefik dynamic config
Snapshot of the two artifacts that control public service exposure as of
2026-07-05. Updated 2026-07-05: fixed Jellyfin backend from dead hubris IP
(192.168.8.206) to new strong IP (192.168.8.246).
## hubris-public-cert-sync.sh (PVE host, `/usr/local/bin/`)
```bash
#!/bin/bash
# Mirrors home caddy's LE certs for publicly-exposed hubris.network hostnames
# into the VPS traefik's /letsencrypt volume. Traefik file-watches the volume
# and hot-reloads.
#
# Why: netbird-proxy's HostSNI(*) TCP passthrough intercepts ACME TLS-ALPN-01
# challenges before traefik's allowACMEByPass can respond, so traefik can't
# obtain its own cert. Home caddy uses IONOS DNS-01 (no such conflict);
# we just mirror what it already has.
#
# Runs daily via hubris-public-cert-sync.timer.
set -euo pipefail
CADDY_LXC=121
CADDY_BASE=/var/lib/caddy/.local/share/caddy/certificates/acme-v02.api.letsencrypt.org-directory
VPS_HOST=root@100.122.165.149
VPS_DEST=/var/lib/docker/volumes/opt_netbird_traefik_letsencrypt/_data
# Map: source hostname -> "crt_filename key_filename" on the VPS.
# Stable names so traefik dynamic.yaml doesn't need edits on renewal.
declare -A HOSTS=(
[artifacto.hubris.network]="fullchain.crt privkey.key"
[blog.hubris.network]="blog.fullchain.crt blog.privkey.key"
[trmnl.hubris.network]="trmnl.fullchain.crt trmnl.privkey.key"
[sso.hubris.network]="sso.fullchain.crt sso.privkey.key"
[media.hubris.network]="media.fullchain.crt media.privkey.key"
[paperless.hubris.network]="paperless.fullchain.crt paperless.privkey.key"
)
TMP=$(mktemp -d)
trap 'rm -rf "$TMP"' EXIT
for host in "${!HOSTS[@]}"; do
read -r crt_name key_name <<< "${HOSTS[$host]}"
pct pull "$CADDY_LXC" "$CADDY_BASE/$host/$host.crt" "$TMP/$crt_name"
pct pull "$CADDY_LXC" "$CADDY_BASE/$host/$host.key" "$TMP/$key_name"
if ssh -o BatchMode=yes "$VPS_HOST" "test -f $VPS_DEST/$crt_name && diff -q - $VPS_DEST/$crt_name" < "$TMP/$crt_name" >/dev/null 2>&1; then
echo "hubris-public-cert-sync: $host unchanged"
continue
fi
scp -q -o BatchMode=yes "$TMP/$crt_name" "$TMP/$key_name" "$VPS_HOST:$VPS_DEST/"
echo "hubris-public-cert-sync: shipped $host ($(openssl x509 -in "$TMP/$crt_name" -noout -enddate))"
done
```
### Adding a new host
1. Caddy must already have the cert (verify `pct exec 121 -- ls "$CADDY_BASE/$host/"`)
2. Add a line to the HOSTS array: `[new-host.hubris.network]="nickname.fullchain.crt nickname.privkey.key"`
3. `systemctl start hubris-public-cert-sync.service` to sync immediately
4. Verify certs landed: `ssh "$VPS_HOST" "ls -la $VPS_DEST/nickname.*"`
5. Add matching `tls.certificates` entry in traefik dynamic config
---
## traefik-dynamic.yaml (VPS, `/opt/`)
```yaml
tcp:
serversTransports:
pp-v2:
proxyProtocol:
version: 2
tls:
certificates:
- certFile: /letsencrypt/fullchain.crt
keyFile: /letsencrypt/privkey.key
- certFile: /letsencrypt/blog.fullchain.crt
keyFile: /letsencrypt/blog.privkey.key
- certFile: /letsencrypt/trmnl.fullchain.crt
keyFile: /letsencrypt/trmnl.privkey.key
- certFile: /letsencrypt/sso.fullchain.crt
keyFile: /letsencrypt/sso.privkey.key
- certFile: /letsencrypt/media.fullchain.crt
keyFile: /letsencrypt/media.privkey.key
- certFile: /letsencrypt/paperless.fullchain.crt
keyFile: /letsencrypt/paperless.privkey.key
http:
routers:
artifacto-public:
rule: 'Host(`artifacto.hubris.network`) && (PathPrefix(`/p/`) || PathPrefix(`/static/`) || Path(`/healthz`))'
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- artifacto-strip-sso
- artifacto-ratelimit
service: artifacto-public
blog-public:
rule: 'Host(`blog.hubris.network`)'
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- blog-ratelimit
service: blog-public
trmnl-public:
rule: 'Host(`trmnl.hubris.network`)'
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- trmnl-ratelimit
service: trmnl-public
matrix-public:
rule: 'Host(`matrix.hubris.network`) && !PathPrefix(`/.well-known/matrix/`)'
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- matrix-ratelimit
service: matrix-public
matrix-wellknown:
rule: 'Host(`matrix.hubris.network`) && (PathPrefix(`/.well-known/matrix/`) || PathPrefix(`/.well-known/acme-challenge/`))'
entryPoints:
- websecure
priority: 20
tls:
certResolver: letsencrypt
service: matrix-wellknown-svc
house-public:
rule: Host(`house.hubris.network`)
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- house-ratelimit
service: house-public
sso-public:
rule: 'Host(`sso.hubris.network`)'
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- sso-ratelimit
service: sso-public
media-public:
rule: 'Host(`media.hubris.network`)'
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- media-ratelimit
service: media-public
paperless-api-public:
rule: 'Host(`paperless.hubris.network`) && PathPrefix(`/api/`)'
entryPoints:
- websecure
priority: 20
tls:
certResolver: letsencrypt
middlewares:
- paperless-ratelimit
service: paperless-public
paperless-public:
rule: 'Host(`paperless.hubris.network`)'
entryPoints:
- websecure
priority: 10
tls:
certResolver: letsencrypt
middlewares:
- authentik-forwardauth
- paperless-ratelimit
service: paperless-public
middlewares:
artifacto-strip-sso:
headers:
customRequestHeaders:
X-Artifacto-Gateway: ""
X-Authentik-Username: ""
X-Authentik-Groups: ""
X-Authentik-Email: ""
X-Authentik-Name: ""
X-Authentik-Uid: ""
X-Authentik-Jwt: ""
X-Authentik-Meta-Jwks: ""
X-Authentik-Meta-Outpost: ""
X-Authentik-Meta-Provider: ""
X-Authentik-Meta-App: ""
X-Authentik-Meta-Version: ""
admin-allowlist:
ipAllowList:
sourceRange:
- "5.61.168.0/24"
artifacto-ratelimit:
rateLimit:
average: 50
period: 1s
burst: 100
blog-ratelimit:
rateLimit:
average: 100
period: 1s
burst: 200
trmnl-ratelimit:
rateLimit:
average: 20
period: 1s
burst: 40
matrix-ratelimit:
rateLimit:
average: 30
period: 1s
burst: 60
house-ratelimit:
rateLimit:
average: 30
period: 1s
burst: 60
sso-ratelimit:
rateLimit:
average: 30
period: 1s
burst: 60
media-ratelimit:
rateLimit:
average: 30
period: 1s
burst: 60
paperless-ratelimit:
rateLimit:
average: 20
period: 1s
burst: 40
authentik-forwardauth:
forwardAuth:
address: "http://192.168.8.6:9000/outpost.goauthentik.io/auth/traefik"
trustForwardHeader: true
authResponseHeaders:
- X-authentik-username
- X-authentik-groups
- X-authentik-email
- X-authentik-name
- X-authentik-uid
- X-authentik-jwt
- X-authentik-meta-jwks
- X-authentik-meta-outpost
- X-authentik-meta-provider
- X-authentik-meta-app
- X-authentik-meta-version
services:
artifacto-public:
loadBalancer:
servers:
- url: 'http://192.168.8.205:3100'
blog-public:
loadBalancer:
servers:
- url: 'http://192.168.8.205:8080'
trmnl-public:
loadBalancer:
servers:
- url: 'http://192.168.8.211:9851'
matrix-public:
loadBalancer:
servers:
- url: 'http://192.168.8.242:8008'
matrix-wellknown-svc:
loadBalancer:
servers:
- url: 'http://matrix-wellknown:80'
house-public:
loadBalancer:
servers:
- url: 'http://192.168.8.244:3000'
sso-public:
loadBalancer:
servers:
- url: 'http://192.168.8.6:9000'
media-public:
loadBalancer:
servers:
- url: 'http://192.168.8.246:8096'
paperless-public:
loadBalancer:
servers:
- url: 'http://192.168.8.130:8000'
```
### Adding a new service — four blocks needed
1. **Router**`http.routers.<name>-public` with `tls: {}` (not
`certResolver`)
2. **Middleware** — rate limit, one per service
3. **Service**`http.services.<name>-public` with the backend URL
4. **tls.certificates** — add a new `- certFile/keryFile` pair matching the
cert sync HOSTS entry
### Key file naming convention
| Cert name | Host | Convention |
|-----------|------|------------|
| `fullchain.crt` + `privkey.key` | `artifacto.hubris.network` | First service — no prefix |
| `blog.fullchain.crt` + `blog.privkey.key` | `blog.hubris.network` | `{nickname}.fullchain.crt` |
| `trmnl.fullchain.crt` + `trmnl.privkey.key` | `trmnl.hubris.network` | `{nickname}.fullchain.crt` |
| `sso.fullchain.crt` + `sso.privkey.key` | `sso.hubris.network` | `{nickname}.fullchain.crt` |
| `media.fullchain.crt` + `media.privkey.key` | `media.hubris.network` | `{nickname}.fullchain.crt` |
| `paperless.fullchain.crt` + `paperless.privkey.key` | `paperless.hubris.network` | `{nickname}.fullchain.crt` |
### ⚠️ Critical — keep backends in sync after LXC migrations
When moving an LXC between Proxmox nodes, update **both**:
1. **Caddy** (`/etc/caddy/Caddyfile` on LXC 121)
2. **VPS traefik** (`/opt/traefik-dynamic.yaml` — via hubris bridge SSH)
Jellyfin migration from hubris to strong (2026-07-05) was fixed in Caddy
but **missed** in VPS traefik — old IP `192.168.8.206` remained. This caused
Bad Gateway for off-LAN users. Use Python-based editing (see
`references/traefik-config-editing.md`) for accurate surgical fixes.